<?xml version="1.0" encoding="UTF-8"?>

<record version="17" id="1505">
 <title>Hermitian matrix</title>
 <name>HermitianMatrix</name>
 <created>2002-01-21 04:18:39</created>
 <modified>2008-10-20 02:20:39</modified>
 <type>Definition</type>
 <creator id="1858" name="matte"/>
 <author id="20947" name="bci1"/>
 <author id="19637" name="invisiblerhino"/>
 <author id="14365" name="Mazzu"/>
 <author id="13753" name="Mathprof"/>
 <author id="1858" name="matte"/>
 <author id="2" name="akrowne"/>
 <classification>
	<category scheme="msc" code="15A57"/>
 </classification>
 <defines>
	<concept>Hermitian operator</concept>
 </defines>
 <synonyms>
	<synonym concept="Hermitian matrix" alias="Hermitian"/>
	<synonym concept="Hermitian matrix" alias="self-adjoint"/>
 </synonyms>
 <related>
	<object name="SelfDual"/>
	<object name="SkewHermitianMatrix"/>
	<object name="SelfAdjointOperator"/>
	<object name="PauliMatrices"/>
 </related>
 <preamble>\usepackage{amssymb}
\usepackage{amsmath}
\usepackage{amsfonts}

%\usepackage{psfrag}
%\usepackage{graphicx}
%\usepackage{xypic}</preamble>
 <content>For a complex matrix $A$, let $A^\ast=\overline{A}^{T}$, where 
$A^T$ is the transpose, and $\bar{A}$ is the complex conjugate of $A$.

{\bf Definition}
A complex square matrix $A$ is \emph{Hermitian}, if 
$$ A = A^*. $$

\subsubsection*{Properties}
\begin{enumerate}
\item The eigenvalues of a Hermitian matrix are real.
\item The diagonal elements of a Hermitian matrix are real.
\item The complex conjugate of a Hermitian matrix is a Hermitian matrix.
\item If $A$ is a Hermitian matrix, and $B$ is a complex matrix
of same order as $A$, then $BAB^\ast$ is a Hermitian matrix.
\item A matrix is symmetric if and only if it is real and Hermitian.
\item Hermitian matrices are a vector subspace of the vector space of 
complex matrices. 
The real symmetric matrices are a subspace of the Hermitian matrices.
\item Hermitian matrices are also called \emph{self-adjoint} since if $A$ is
Hermitian, then in the usual
inner product of $\mathbb{C}^n$, we have 
$$ \langle u,Av \rangle = \langle Au,v\rangle$$
for all $u,v\in \mathbb{C}^n$.

\end{enumerate}

\subsubsection*{Example}
\begin{enumerate}
\item For any $n\times m$ matrix $A$, the $n\times n$ matrix $A A^\ast$ is
Hermitian. 
\item For any square matrix $A$, the Hermitian part of $A$,  
      $\frac{1}{2}(A+A^\ast)$ is Hermitian. 
      See \PMlinkname{this page}{DirectSumOfHermitianAndSkewHermitianMatrices}.
\item 
$$ \begin{bmatrix}
  1 &amp; 1 + i &amp; 1 + 2i &amp; 1 + 3i \\
  1 - i &amp; 2 &amp; 2 + 2i &amp; 2 + 3i \\
  1 - 2i &amp; 2 - 2i &amp; 3 &amp; 3 + 3i \\
  1 - 3i &amp; 2 - 3i &amp; 3 - 3i &amp; 4
\end{bmatrix} $$
\end{enumerate}
The first two examples are also examples of normal matrices.

\subsubsection*{Notes}
\begin{enumerate}
\item Hermitian matrices are named after Charles Hermite (1822-1901) \cite{hermite}, who proved  in 1855 that the 
eigenvalues of these matrices are always real \cite{eves}.
\item Hermitian, or self-adjoint operators on a Hilbert space play a fundamental
role in quantum theories as their eigenvalues are observable, or measurable; such
Hermitian operators can be represented by Hermitian matrices.
\end{enumerate}

\begin{thebibliography}{9}
 \bibitem{eves} H. Eves,
 \emph{Elementary Matrix Theory},
 Dover publications, 1980.
\bibitem{hermite}
 The MacTutor History of Mathematics archive, 
\PMlinkexternal{Charles Hermite}{http://www-gap.dcs.st-and.ac.uk/~history/Mathematicians/Hermite.html}
\end{thebibliography}</content>
</record>
